Going to Megasquirt, More questions Options

[bbrock]

After about 1,700 miles of driving my freshly restored car with rebuilt engine, I’ve decided I’ve had my fun with Weber carbs and distributor-based ignition and am ready to step up to modern EFI and coil on plug ignition. Even with the jetting that came out of the box leaving the carbs running rich, the performance has been fantastic. What is not fantastic is the garage stinking of gasoline, no compensation for altitude, and being generally too fiddly to set up for my taste. My trials with the ignition are documented in another thread. Yes, I know a 1-2-3 would solve those problems but for the money, I’d rather invest in modern COP ignition as part of an EFI upgrade.

My goal is a smooth, efficient, and reliable street machine that I can drive from sea level to 11,000 ft. without starving or choking on fuel. Efficiency is at least as important as performance. As long as I can get at least the stockish 100 hp, I’ll be happy and beyond that, I’d like to wring as many mpg out as possible.

The engine is a mostly stock euro-spec 2L engine. The only mod is a fairly mild Elgin 6048 camshaft with 256 duration for the carbs. A source of pride of this build is this custom 911/914-6 inspired air cleaner I made which I think looks cool and really silences the carbs.



Now for the questions:

• Single throttle body or ITB? I think I’ve made a decision but still interested in thoughts. I was thinking about welding injector bungs onto the carb manifolds and using my carbs as throttle bodies. The main appeal is that I would keep my cool air cleaner to make the other kids jealous. However, it seems the stock throttle body would greatly simplify the conversion. Also, even though the custom intake is designed to allow access for servicing and easy air filter replacement, it does crowd an already crowded engine bay and makes working in there just that much more of a challenge. My stock TB needs some TLC and might have to be sent for professional refurbishing. I think I could sell my carb setup to cover that cost but not sure. The upshot is that I’ve all but decided to go back to the stock TB, but curious what others think.

• N Alpha, Speed Density, or MAF? I’ve been reading up on this and think I understand pros and cons, but still a little confused about sensors needed. With my efficiency goal, I think MAF is the way to go. It looks to me that cutting off the tube connecting the stock air cleaner to the TB and replacing it with a MAF could be a really slick way to add MAF in stealth fashion. Has anyone done this? If not, how does one find the right MAF to use? Other than dimensions, what else needs to be considered?

Another question is about MAP + MAF vs MAF only. I’m a little confused about advantages or when a MAP sensor is needed if you have a MAF.

• Barometric correction – this is an important feature for my location, but the hardware needed to implement it is a little confusing. It seems like if you are running a MAP, then barometric correction is obtained by adding a second pressure sensor (another MAP?) to read reference atmospheric pressure to make corrections to the fuel mixture. How does it work with MAF? Do you only need one pressure sensor to read atmosphere? Or do you still need to reference it against manifold pressure? I assume a lot of this is done in the software but I haven’t looked to far into the tuning part yet. I’m more trying to figure out a shopping list for parts at this point.

• Anyone running a CAM sync and sequential spark and injection? Again with the efficiency goal, this is appealing. Looks like Mario is working on a new version which isn’t available yet, are there alternatives available? It seems people say you still need a crank position sensor even with a cam sensor in the mix. It isn’t entirely clear why though. Lastly, and this is mostly just curiosity, but is it correct to think that the lifespan (in miles) of spark plugs are cut in half with wasted spark?

I have many more questions but this is already too long so will save them for later.
TIA

[GregAmy]

$12 to add voltage input for CHT to my Microsquirt for monitoring and logging head temps.

https://www.adafruit.com/product/1778

This thread has gone off the rails, from driving a 50-yr-old car (with a 75-yr-old engine design) in The Real World to thinking this can be micro-controlled to a modern level of efficiency. (IMG:style_emoticons/default/wink.gif)

Just build something already!

And if you do follow through on this science project, it would be very cool if you did it in such a way that it could be re-tuned to basic wasted spark and batch injection, and you spent the dyno time to prove this was worth all those extra brain cycles and all that extra financial outlay. I know which spot on the table my twenty-spot would be palced on. (IMG:style_emoticons/default/wink.gif)

Have fun!

GA

[jd74914]

That’s not an uncommon feature in higher end ECUs. I wonder if you could bake it into a corrections table. The thing I’m not sure about it usually those cascade control loops have a large time constant between them (ie coolant temp correction runs 100x slower than O2 feedback). Of course, if the change is really small that likely doesn’t matter much.

Reading the TC in would be easy with an amplifier. You can get cheap ones on Sparkfun.

[Frank S]

Still not sure how much precision is being sacrificed using a gear driven distributor to send tach signal vs. a crank sensor.

This is entirely observational but "some"

I went from picking up my tach signal from a pertronix on a locked dizzy to Marios 36 tooth crank wheel and the difference under a timing light was visibly noticeable. couldn't tell you in a number what the difference was and im sure its very minor, like less than 1 deg, but with the dizzy trigger there was a bit of jitter under the timing light, the timing mark had a slight haze to it. After moving to the 36 tooth crank wheel the mark under a timing light was rock solid, like i was looking at a laser pointer.

I think thats the jitter I'm seeing as well. But does that really mater?
So many here report back here that Microsquirt is working well for them and Microsquirt as well as MS2 have random injection timing (different from engine start to engine start and not paired to a specific event) if you run them batch injected. This is the main reason why I changed from MS2 to MS3 as the code for Injection timing is totaly different and also supports Timed Batch injection as the D-Jet does.
With this in place, the engine is behaving the same from egine start to engine start (clearly noticable if you compare the datalogs).
Also for ignition timing you are typically some degrees away from where spark knock really happens. So the 1° Jitter should be also fine therefore.

[Frank S]

$12 to add voltage input for CHT to my Microsquirt for monitoring and logging head temps.

https://www.adafruit.com/product/1778

This thread has gone off the rails, from driving a 50-yr-old car (with a 75-yr-old engine design) in The Real World to thinking this can be micro-controlled to a modern level of efficiency. (IMG:style_emoticons/default/wink.gif)

Just build something already!

And if you do follow through on this science project, it would be very cool if you did it in such a way that it could be re-tuned to basic wasted spark and batch injection, and you spent the dyno time to prove this was worth all those extra brain cycles and all that extra financial outlay. I know which spot on the table my twenty-spot would be palced on. (IMG:style_emoticons/default/wink.gif)

Have fun!

GA

Sequential injection does not generate more HP than batch injection. But you can lean it out much better in partload which is leading to better fuel economy. Thats all.

[bbrock]

Still not sure how much precision is being sacrificed using a gear driven distributor to send tach signal vs. a crank sensor.

This is entirely observational but "some"

I went from picking up my tach signal from a pertronix on a locked dizzy to Marios 36 tooth crank wheel and the difference under a timing light was visibly noticeable. couldn't tell you in a number what the difference was and im sure its very minor, like less than 1 deg, but with the dizzy trigger there was a bit of jitter under the timing light, the timing mark had a slight haze to it. After moving to the 36 tooth crank wheel the mark under a timing light was rock solid, like i was looking at a laser pointer.

I think thats the jitter I'm seeing as well. But does that really mater?
So many here report back here that Microsquirt is working well for them and Microsquirt as well as MS2 have random injection timing (different from engine start to engine start and not paired to a specific event) if you run them batch injected. This is the main reason why I changed from MS2 to MS3 as the code for Injection timing is totaly different and also supports Timed Batch injection as the D-Jet does.
With this in place, the engine is behaving the same from egine start to engine start (clearly noticable if you compare the datalogs).
Also for ignition timing you are typically some degrees away from where spark knock really happens. So the 1° Jitter should be also fine therefore.

One degree seems in line with my very unscientific method of wiggling the dizzy rotor to see how much play it has.

[Montreal914]

For the CHT sensor, I will be using a BMW motorcycle sensor 12627673407. It has the same thread as the stock D-Jet CHT. You can see it here with the #3 ring sensor that will be used as a readout. I might also use Greg Amy's gizmo to datalog.




The CHT installation tool I did using a modified FLAP long reach socket.




Last pictures is my fancy (IMG:style_emoticons/default/rolleyes.gif) calibration process.







My microsquirt conversion will resume (IMG:style_emoticons/default/headbang.gif) once I send out my shell to the paint shop... Hopefully in the next few months (IMG:style_emoticons/default/rolleyes.gif)

[bbrock]

This thread has gone off the rails, from driving a 50-yr-old car (with a 75-yr-old engine design) in The Real World to thinking this can be micro-controlled to a modern level of efficiency. (IMG:style_emoticons/default/wink.gif)

Oh good. I'm in my comfort zone then (IMG:style_emoticons/default/biggrin.gif)

Thanks for the link to the amplifier. I remembered you had done that and was going to look up the thread where you described it to bookmark (IMG:style_emoticons/default/thumb3d.gif)

I'll reveal my ignorance, but can a dyno tell us about differences in fuel economy? As Frank said, that's the only place I expect to see any difference. It seems advances in fuel and spark management are major factors in efficiency improvement of modern engines. Granted it is old school intake and combustion, and cooling technology so there will be limits.

As for performance - I've always found the stock 2.0L almost perfectly suited for my needs.

So continuing the theme of off the rails slippery slope, looks like I'll need some sort of gps integration for vehicle speed to be able to log mpg.

I'm going to order the MS3 kit soon but this project realistically won't begin in earnest until at least mid-December. It doesn't matter anyway since I want to buy a fair amount of kit from Mario and can't do that until he opens for shopping again.

[bbrock]

For the CHT sensor, I will be using a BMW motorcycle sensor 12627673407. It has the same thread as the stock D-Jet CHT. You can see it here with the #3 ring sensor that will be used as a readout. I might also use Greg Amy's gizmo to datalog.

That looks like a great solution! I was going to use the modded GM sensor from the Dub Shop that bolts to the stock CHT location with an M10 bolt. I already have the bolt in there just to prevent any gunk from getting in the hole. Will still need to source a ring sensor #3 plug.

[Montreal914]

Will still need to source a ring sensor #3 plug.

This is an option: K-type TC Rings, readers w/outputs, etc...

http://thesensorconnection.com/

[Superhawk996]

can a dyno tell us about differences in fuel economy?

Not really. At least with old school water brake dyno's and eddy current dyno's. They don't handle transients very well nor can they simulate engine coastdowns when you are off throttle.

EPA fuel economy cycle is driven on a elaborate chassis dyno so as to include all parasitic loads. There is a ton of vehicle level coast down work required to establish the proper load parameters to feed to the EPA chassis dyno so that it can mimic proper aerodynamic and road loads. Real world fuel economy isn't going to be easily captured on an engine dyno.

An engine dyno can tell you about fuel consumption at a given load. That's about it.
https://en.wikipedia.org/wiki/Brake-specifi...uel_consumption

This is of course a bit oversimplied -- there are newer AC dynos that can run transients but that isn't what you're going to find with Joe's Hot Rod shop of wherever you might rent dyno time. If you want more info Google AVL dynos, Horiba dynos, or FEV consulting to get an idea what's out there with the big boys of the dyno world that support industry OEM's.

[ClayPerrine]

The CHT sensor for a 911 Carrera 3.2 is the same as the 914 one, but it is a two wire sensor so it doesn't depend on the engine case for ground.

Same curve as a 914 sensor, so it is easy to make work with Megasquirt and fits the cylinder head perfectly.

Clay

[worn]

. After moving to the 36 tooth crank wheel the mark under a timing light was rock solid, like i was looking at a laser pointer.

When I was in graduate school I had to give a seminar talk each year. After my first presentation before a live audience a friend pointed out that a laser pointer exaggerates how nervous you are. A bit of fear mixed with caffeine will send that laser point all over creation. Let’s just say I was not ever rock solid.